Enhancing Physical-Layer Security via Big-Data-Aided Hybrid Relay Selection

The explosive growth in data traffic presents new challenges to the new generation of wireless communication systems, such as computing capabilities, spectrum efficiency and security. In this paper, we use the network structure, which is adaptable for the big data traffic, to improve the security of wireless networks. Specifically, a big-data aided hybrid relay selection scheme is designed and analyzed to enhance physical layer security. First, considering the ideal situation that an eavesdropper's CSI (Channel State Information) is known to the legal nodes, we propose an optimal hybrid relay selection scheme consisting of the optimal mode selection scheme and the optimal relay selection scheme. In this case, we analyze the upper bound of an eavesdropper's capacity in FD (Full-Duplex) mode and the secrecy outage probabilities of the optimal HD (Half-Duplex), FD, and hybrid relay selection schemes. Through the analysis of data, it is clear that the mode selection is decided by the self-interference of the FD technique. However, the instantaneous CSI of an eavesdropper is difficult to obtain due to the passive characteristic of eavesdroppers in practice. Therefore, a more practical hybrid relay selection scheme with only the channel distribution information of an eavesdropper is further studied, where a weighting factor is employed to guarantee that the hybrid mode is no worse than either the FD mode or HD mode when the self-interference grows. Finally, the simulation results show the improved security of our proposed scheme.